Implementing integrated genomic risk assessments for breast cancer: lessons learned from the Electronic Medical Records and Genomics study

Abstract

Objectives: To implementation an automated multi-institutional pipeline that delivers breast-cancer risk integrated with polygenic risk scores, monogenic variants, family history, and clinical factors, emphasizing operational challenges and their solutions.

Materials and methods: A five-stage process was executed at ten sites. Data streams from REDCap surveys, PRS and monogenic reports, and MeTree pedigrees were normalized and forwarded through a REDCap plug-in to the CanRisk API.

Results: Integrated risk was returned to >10 000 women; 3.6% were ≥25 % lifetime risk and 0.9% carried pathogenic variants. Pipeline generated score aligns well with manual generated ones. Major barriers such as heterogeneous pedigree formats, missing data, edge-case handling, and evolving model versions were identified and resolved through mapping rules, imputations, and iterative testing.

Discussion: Cross-platform data harmonization and stakeholder alignment were decisive for success. Borderline-risk communication and model-version drift remain open issues.

Conclusion: Large-scale PRS-integrated breast-cancer risk reporting is feasible but requires robust interoperability standards and iterative governance.

Keywords: breast cancer; clinical workflow automation; integrated risk score; polygenic risk score.

Figure 1.

Swimlane diagram illustrating the overview of stakeholders and implementation activities. Different colors/lanes represent distinct stakeholders. Each box indicates a specific task involved in the overall implementation process. Dashed boxes denote iterative processes.

Figure 2.

Finalized standard operating procedure (SOP) for the risk generation, review, validation, and return process. The entire procedure is composed of multiple phases (different colors) and includes five checkpoints (diamond-shaped boxes). Phase 1: Before API query, the participant must have sex assigned as female at birth, 18 years and old, and provided baseline height and weight. If MeTree data are unavailable, the site’s research staff acknowledged its absence, and in such cases, the R⁴ extension generated a placeholder pedigree containing only healthy parents (father and mother). Other missing data (eg, PRS) were permissible. For a successful API risk generation, the participant must have no prior breast cancer history recorded in the MeTree, and no monozygotic twins; Phase 2: Borderline cases were review. Sites then verified that data from Broad, Invitae, MeTree, and other sources was reviewed and approved before generating the GIRA report, which includes conditions beyond breast cancer risk. Phase 3: Sites implemented their own high-risk return protocols to communicate risk to participants and their healthcare providers. If additional information (eg, from the EHR) revealed a prior history of breast cancer, sites were instructed not to return a high-risk breast cancer result.

Figure 3.

Distribution of life-time risk and PRS in different self-reported race/ethnicity groups. Breakdown statistics were listed only four groups with more than 500 integrated score generated. Individuals might report more than one race/ethnicity. “Overall” includes other ancestries not listed in the breakdown categories. Both Invitae report and PRS could be missing in generating BOADICEA. High PRS/Monogic risk ratio is calculated among PRS/Invitae reports available individuals.